Originating portfolio: NET
Estimated expiration: 2024-Jan-22
Potentially relevant companies (2): Amazon, Google Inc.
Products & technologies (2): Audio:Processing, Speech:Processing
A method of detecting voice activity in a digitally coded speech signal propagating in uplink direction for a mobile communication system , said coded speech signal comprising a representation of a spectral information and received at a receiver (4, 6) belonging to a control part (1, 3) of said system, in which said detection is performed by operating directly on the received speech signal and a signal representative of the presence or absence of voice activity is generated by submitting the speech signal to an energetic analysis, characterised in that said detection also includes a spectral analysis performed directly on the received coded speech signal, by exploiting only the spectral information contained in the received signal itself; and spectrum and energy information resulting from said spectral analysis and said energetic analysis, respectively, are jointly processed for generating said signal representative of the presence or absence of voice.
Interested in licensing a slice of this patent? Contact us to take the next step, or read about our method to understand the logistics.